Ephemeral Storage
Ephemeral Storage ([ɪˈfemərəl ˈstɔːrɪʤ]): a storage without strong durability (сохранность) guarantees. Keeps data that can be infrequently (~once per day) lost. Possible implementation: LRU Cache that is lost after a VVM restart.
Motivation
github: Ephemeral Storage #2269
674859
As a location owner, I want to prevent POS users from working on the same table simultaneously to avoid confusion and errors. We want to reduce the likelihood of such errors occurring.
Use cases
Lock the table (restaurant) so that waiters can not work with the same table simultaneously.
Extra measures shall be undertaken to prevent the conflict ("this table has been already modified by someone else").
Key: {PartitionID, WSID, subject(device), tableNo}, Value: {lockedAt, userID}.
Track the status of devices.
Key: {PartitionID, WSID(app workspace), subject(device), }, Value: {lastActivity}.
Keep demo operation data (
Cloud Bills
).
Functional design
Principles
LRU Cache with fixed-size entries, that is lost after a VVM restart.
By default 5MB per partition.
Fixed-size entries => keys and values have basic types.
Principles
VQL:
STORAGE Ephemeral(
GET SCOPE(COMMANDS, QUERIES),
INSERT SCOPE(COMMANDS),
UPDATE SCOPE(COMMANDS),
READ SCOPE(QUERIES),
);
Deployment descriptor:
???
Last updated
Was this helpful?